western interconnection flexibility assessment · 2016-01-14 · 101 montgomery street, suite 1600...

Western Interconnection Flexibility Assessment

Executive Summary

December 2015

© 2015 Energy and Environmental Economics, Inc.

© 2015 Copyright. All Rights Reserved.

Energy and Environmental Economics, Inc.

101 Montgomery Street, Suite 1600

San Francisco, CA 94104

415.391.5100

www.ethree.com

Project Team:

Nick Schlag, Arne Olson, Elaine Hart, Ana Mileva, Ryan Jones (E3)

Carlo Brancucci Martinez‐Anido, Bri‐Mathias Hodge, Greg Brinkman, Anthony Florita,

David Biagioni (NREL)


Executive Summary

December 2015

P a g e | i |© 2015 Energy and Environmental Economics, Inc.

About this Study

This study was jointly undertaken by the Western Electricity Coordinating

Council (WECC) and the Western Interstate Energy Board (WIEB) to investigate

the need for power system flexibility to ensure reliable and economic

operations of the interconnected Western electricity system under higher

penetrations of variable energy resources. WECC and WIEB have partnered with

Energy & Environmental Economics, Inc. (E3) and the National Renewable

Energy Laboratory (NREL) to investigate these questions using advanced,

stochastic reliability modeling and production cost modeling techniques. The

study identifies and examines operational challenges and potential enabling

strategies for renewable integration under a wide range of operating conditions,

scenarios, and sensitivities across the Western Interconnection, with the goal of

providing guidance to operators, planners, regulators and policymakers about

changing system conditions under higher renewable penetration.

Funding for this project is provided by a number of sources. E3’s role in the

project is funded jointly by WECC and WIEB through grants received under the

Department of Energy’s American Recovery and Reinvestment Act (ARRA);

NREL’s role is funded directly by the Department of Energy.

Technical Review Committee

This study was overseen by a Technical Review Committee (TRC) comprising

representatives from utilities, regulatory agencies, and industry throughout the

Western Interconnection:

P a g e | ii |

Aidan Tuohy, Electric Power Research Institute

Ben Kujala, Northwest Power Planning & Conservation Council

Brian Parsons, Western Grid Group

Dan Beckstead, Western Electricity Coordinating Council

Fred Heutte, Northwest Energy Coalition

James Barner & Bingbing Zhang, Los Angeles Department of Water &

Power

Jim Baak, Vote Solar Initiative

Justin Thompson, Arizona Public Service

Keith White, California Public Utilities Commission

Michael Evans, Shell Energy North America

Thomas Carr, Western Interstate Energy Board

Thomas Edmunds, Lawrence Livermore National Laboratory

Tom Miller, Pacific Gas & Electric Company

The TRC met a number of times throughout the course of the project to provide

input and guidance on technical modeling decisions, to help interpret analysis,

and to craft the study’s ultimate findings and conclusions. The feedback of the

TRC throughout the project was highly valuable to the project.

Executive Summary

P a g e | iii |© 2015 Energy and Environmental Economics, Inc.

Executive Summary

Background

Over the past decade, the penetration of renewable generation in the Western

Interconnection has grown rapidly: nearly 30,000 MW of renewable generation

capacity—mostly solar and wind—have been built. By 2024, under current state

policies, the total installed capacity of renewable generation in the Western

Interconnection may exceed 60,000 MW. In front of this landscape of increasing

renewable policy targets and declining renewable costs, interest in renewable

generation and understanding its impacts on electric systems has surged

recently. Regulators, utilities, and policymakers have begun to grapple with the

potential need for power system “flexibility” to ensure reliable operations under

high penetrations of renewable generation.

The Western Electricity Coordinating Council (WECC), in its role as the Regional

Entity responsible for reliability in the Western Interconnection, is interested in

understanding the long‐term adequacy of the interconnected western grid to

meet the new operational challenges posed by wind and solar generation across

a range of plausible levels of penetration. WECC stakeholders have expressed a

similar interest through study requests to examine high renewable futures that

that implicate operational and flexibility concerns. The Western Interstate

Energy Board (WIEB) seeks to understand these issues in order to inform


P a g e | iv |

policymakers about the implications of potential future policies targeting higher

renewable penetrations.

With this motivation, WECC and WIEB collaborated to jointly sponsor this WECC

Flexibility Assessment. The sponsors established three goals for this effort:

Assess the ability of the fleet of resources in the Western

Interconnection to accommodate high renewable penetrations while

maintaining reliable operations. Higher penetrations of renewable

generation will test the flexibility of the electric systems of the West by

requiring individual power plants to operate in fundamentally new

ways, changing operating practices as well as the dynamics of wholesale

power markets. This study aims to identify the major changes in

operational patterns that may occur at such high penetrations and to

measure the magnitude and frequency of possible challenges that may

result.

Investigate potential enabling strategies to facilitate renewable

integration that consider both institutional and physical constraints on

the Western system. Existing literature has identified a wide range of

possible strategies that may facilitate the integration of high

penetrations of renewables into the Western Interconnection. These

strategies comprise both institutional changes—for example, increased

use of curtailment as an operational strategy and greater regional

coordination in planning and operations—as well as physical changes to

the electric system—new investments in flexible generating resources

and the development of new demand side programs. This study

examines how such measures can mitigate the challenges that arise

with interconnection‐wide increases in renewable penetration.

Provide lessons for future study of system flexibility on the relative

importance of various considerations in planning exercises. The study

Executive Summary

P a g e | v |© 2015 Energy and Environmental Economics, Inc.

of flexibility and its need at high renewable penetrations is an evolving

field. This effort is designed with an explicit goal of providing useful

information to modelers and technical analysts to improve analytical

capabilities for further investigation into the topics explored herein.

Assessment Methods

This study is organized into two sequential phases of analysis. The first phase, a

resource adequacy assessment of the generation fleet, uses traditional loss‐of‐

load‐probability techniques to ensure that the electric system adheres to a

traditional “one‐day‐in‐ten‐year” planning standard for loss of load. The second

phase, the flexibility assessment, uses stochastic production cost analysis to

examine the degree to which operational challenges are encountered with the

additional of renewables to the system. By nesting the flexibility assessment

within a traditional study of resource adequacy, the approach used in this study

seeks to ensure that any challenges encountered in operations can be attributed

to a lack of flexibility and are not simply the result of a system whose available

capacity is inadequate to meet its peak demands.

The first phase of the analysis uses E3’s Renewable Energy Capacity Planning

(RECAP) model, a loss‐of‐load probability (LOLP) model designed to evaluate

resource adequacy under high renewable penetrations. The RECAP analysis

confirmed that the modeled resources across the study area were capable of

meeting or exceeding the traditional one‐in‐ten loss of load frequency


P a g e | vi |

standard.1 The second phase uses the Renewable Energy Flexibility (REFLEX)

model for PLEXOS for Power Systems—an adaptation of traditional production

cost modeling that incorporates principles from more traditional reliability

analysis—to examine the impacts of high penetrations of renewable generation

on the operations of these resources. Production simulation models are used

for a variety of purposes, but the approach taken in this study has been tailored

directly to the examination of flexibility challenges under high renewable

penetrations.

While loss‐of‐load probability modeling is the de facto standard method for

assessing conventional capacity adequacy, there is no analogous industry‐

standard approach for assessing the adequacy of flexibility in an electric system.

A number of approaches have been explored, and a variety of metrics have

been proposed as a means of measuring flexibility adequacy—for instance,

“Expected Unserved Ramp.”2 While these metrics may be useful as indicators of

power system inflexibility, they are not directly actionable because there is no

standard for unserved ramping that power systems are required to meet.

Rather than attempting to define a new metric, the REFLEX production cost

modeling approach used in this study is designed to measure the consequences

of inadequate flexibility, thereby illustrating a method through which actionable

information can be provided to planners and decision‐makers.

1 The results of the Phase 1 analysis, summarized in the body of this report, were also published as a standalone report available at: http://westernenergyboard.org/wp‐content/uploads/2015/06/04‐2015‐WECC‐WIEB‐Flexibility‐Assessment‐Report‐Interim.pdf 2 See, for example, EPRI’s “Power System Flexibility Metrics: Framework, Software Tool and Case Study for Considering Power System Flexibility in Planning,” available at: http://www.epri.com/abstracts/Pages/ProductAbstract.aspx?productId=000000003002000331.

Executive Summary

P a g e | vii |© 2015 Energy and Environmental Economics, Inc.

In traditional LOLP analysis, the consequence of inadequate capacity is

straightforward: the system is incapable of simultaneously meeting all loads, so

some loads cannot be served. This is measured using conventional metrics such

as LOLP, Loss of Load Expectation (LOLE), and Expected Unserved Energy (EUE).

Inadequate flexibility can also lead to loss of load, even for systems that would

otherwise be resource adequate. For example, if a large portion of the thermal

fleet is cycled off to accommodate high output from renewable generators, the

system might not have adequate resources committed to meet a large upward

net load ramp (this is illustrated in Figure 1a). REFLEX therefore tracks EUE as

one measure of power system inflexibility.

However, loss of load can be avoided through prospective curtailment of

renewable generation to ensure that the thermal resources required to

maintain reliability can remain online. This is illustrated in Figure 1b, where

renewable generation is curtailed during the mid‐afternoon in order to ensure

that the system has sufficient operating capability to meet the evening ramp. In

this way renewable curtailment is both a key operational strategy in flexibility‐

constrained systems and is also the primary consequence of inadequate

flexibility.

Because the system operator has a choice between curtailing renewables and

curtailing loads, it needs a method for determining which strategy to use on a

given day. Additional information is therefore needed about the economic

consequences of unserved energy and curtailed renewables. This information

may come through market‐based bids provided by loads and renewable project

operators. In the absence of market information, deemed values are used based

on available literature.


P a g e | viii |

Studies typically, ascribe a high value of between $10,000 and $100,000/MWh

of lost load. In contrast, the cost of renewable curtailment is effectively smaller

by orders of magnitude. Out‐of‐pocket costs of curtailment include O&M costs

as well as any lost production tax credits. In addition, under a production quota

such as an RPS, curtailed renewable generation must typically be replaced like‐

for‐like in order to ensure compliance. Thus, the cost of curtailment is

determined by the “replacement cost” of renewable generation (the cost of

procuring an additional MWh of generation to “replace” the curtailed energy).

In this study, the penalty prices assumed for loss of load and renewable

curtailment are $50,000 and $100/MWh, respectively. The asymmetry between

these penalties ensures that REFLEX will steer the system away from loss of load

if at all possible, allowing some amount of curtailment when necessary to

ensure that the system has adequate operating flexibility. In this respect,

renewable curtailment serves as the “default solution” for system operators

facing flexibility constraints, the relief valve with which an operator can ensure

the electric system remains within an operable range.

Executive Summary

P a g e | ix |© 2015 Energy and Environmental Economics, Inc.

Figure 1. Tradeoff between upward and downward flexibility challenges

In order to capture an adequately broad sample of operating conditions and

possible flexibility constraints using this framework, this study uses Monte Carlo

sampling of load, wind, solar, and hydro across a wide range of historical


P a g e | x |

conditions. Production cost model are typically used to analyze a single “typical”

year with its unique set of load, wind, solar, and hydro profiles. This study

makes use of much longer historical records for each of these variables, as

summarized in Figure 2. Inputs for these variables are derived from a variety of

sources: load shapes based on historical weather conditions spanning a thirty‐

year period are created using a neural network regression model; wind and

solar profiles are derived from NREL’s WIND and SIND Toolkits, respectively; and

hydro data is based on a combination of actual and simulated historical data for

Western hydro systems obtained from the Energy Information Administration

(EIA) and the Northwest Power and Conservation Council (NWPCC). Profiles

from different periods are combined using a stratified sampling methodology to

produce Monte Carlo “day draws” for production cost modeling. By adapting a

technique that has historically been used in LOLP analyses, this approach

captures a more robust distribution of expected long‐run conditions than any

single historical year.

Figure 2. Historical conditions incorporated into draws

2007

2007

1980

1970

2012

2008

2012

2012

1970 1980 1990 2000 2010

Load

Wind

Solar

Hydro

Executive Summary

P a g e | xi |© 2015 Energy and Environmental Economics, Inc.

Figure 3. Regions included in analysis.

The analytical framework described above is used to evaluate the flexibility of

five regions within Western Interconnection, shown in Figure 3. Each region

represents a portion of the Western Interconnection with relatively

homogeneous loads and resources, some existing degree of coordination in

resource planning and/or operations, and limited internal transmission

constraints.3 In the study, each region is linked to its neighbors with a zonal

transmission model. One of the key simplifying assumptions of this approach is

that each region is perfectly coordinated in operations internally, effectively

pools all of its generation resources to balance net load without regard for

existing contracts, ownership, or operational conventions.

3 The Alberta Electric System Operator (AESO) and BC Hydro (BCH) are excluded from this analysis due to lack of data availability; their exclusion from the study may overstate the magnitude of flexibility challenges, particularly if the flexibility of the hydro system in British Columbia could be used to facilitate renewable integration


P a g e | xii |

This study begins with a conservative assumption that exchange between any

two regions is limited to the range that has been observed historically rather

than the full physical capability of the transmission system that links them. This

approach serves three purposes: (1) it serves as a proxy for the many

institutional constraints that exist in today’s system, in which power exchange

between regions generally occurs on a limited and bilateral basis; (2) it helps to

isolate the renewable integration challenge in each region in order to

characterize each fleet’s ability to integrate its own renewable portfolio rather

than relying on the flexibility and diversity of the entire Western system; and (3)

it provides a useful counterfactual to a scenario in which limits on the

transmission system are relaxed to their full physical capability, allowing this

study to highlight the value that could be achieved through centralized dispatch

of all resources in the WECC region.4

Two renewable portfolios are examined within the flexibility assessment for

each region. The first, shown in Figure 4a, is based on the 2024 Common Case

developed by WECC’s Transmission Expansion Planning Policy Committee

(TEPPC) and represents a penetration of renewable generation that is largely

consistent with state RPS targets current as of 2014;5 analysis of this case

4 Increased regional coordination is one of the integration “solutions” examined in this report, and this contrast provides the underlying method through which its value is characterized. 5 At the time of the development of the 2024 Common Case database and the start of this study, California’s RPS policy remained at 33%. With the increase to a 50% target by 2030, the Common Case is no longer indicative of current policy in that region but still provides a useful point of reference for its near‐term relevance.

Executive Summary

P a g e | xiii |© 2015 Energy and Environmental Economics, Inc.

provides a means of validating the model as well as an indication of potential

flexibility challenges that could emerge under existing policy. The second

portfolio studied is a “High Renewables Case,” whose composition is shown in

Figure 4b. The penetrations chosen for study in the High Renewables Case were

intentionally chosen to be high enough to cause significant changes in

operations, and to ensure that flexibility challenges would result, allowing this

study to characterize renewable integration challenges that may emerge above

current policy levels.

Figure 4. Renewable portfolios analyzed in the flexibility assessment


P a g e | xiv |

Nature of Regional Renewable Integration Challenges

Because the penalty prices for unserved energy and curtailment prioritize load

service over the delivery of renewable generation, renewable curtailment is the

key indicator of a system that is constrained in its ability to integrate

renewables. Curtailment can occur due to simple oversupply—where the

generation available exceeds the load—or as a means to ensure reliable

operations in the presence of dispatch flexibility constraints. REFLEX produces

many metrics that are useful for understanding how high penetrations of

renewables impact a system; however, this study uses renewable curtailment as

the key indicator of flexibility constraints for each case.

Table 1 summarizes the renewable curtailment observed in each region for the

renewable portfolios examined. In the Common Case, the extent of renewable

curtailment experienced across the Western Interconnection is limited—it

constitutes less than 0.1% of the available renewable energy. In contrast, in the

High Renewables Case, renewable curtailment appears routinely and represents

a large share of the available renewable generation. The specific nature of the

challenges differ across regions—each a unique result of the region’s distinct

renewable portfolio, the characteristics of its conventional generators, and the

profile of its loads.

Executive Summary

P a g e | xv |© 2015 Energy and Environmental Economics, Inc.

Table 1. Renewable curtailment observed in the Common Case and High Renewables Case (% of annual generation)

Scenario Basin California Northwest Rockies Southwest WECC‐US

Common Case 0.01% 0.00% 0.06% 0.13% 0.00% 0.02%

High Ren Case 0.4% 8.7% 5.6% 0.6% 7.3% 6.4%

In the High Renewables Case, renewable curtailment is most prevalent in

California and is driven by a phenomenon that has been well‐characterized in

past studies of renewable integration: because solar PV resources produce only

during daylight hours, a large share of the generation in California’s renewable

portfolio is concentrated in the middle of the day. The high daytime renewable

production exerts pressure on the non‐renewable fleet to reduce its output to

low levels, but non‐renewable generation is limited in its ability to do so by

inflexibility (especially for must‐run resources like nuclear & cogeneration),

minimum generation constraints, and the need to carry contingency and

flexibility reserves. The result of this dynamic is a regular daily pattern of midday

renewable curtailment illustrated in Figure 5, which shows a typical spring day

in California in the High Renewables Case.


P a g e | xvi |

Figure 5. Typical spring day, California, High Renewables Case6

The Southwest also experiences a large volume of renewable curtailment in the

High Renewables Case. Much like California, the Southwest relies heavily on

solar PV resources in the High Renewables Case and, as a result, experiences a

similar diurnal oversupply phenomenon. Figure 6 presents a typical spring day in

the Southwest in which many of the same elements that characterize the

California challenge are apparent: large diurnal net load ramps that coincide

with sunrise and sunset and a mid‐day period of renewable curtailment when all

dispatchable resources have been reduced to their lowest output.

6 Note that this study reports all results in Pacific Standard Time; all plots that show hourly results use the PST convention.

Executive Summary

P a g e | xvii |© 2015 Energy and Environmental Economics, Inc.

Figure 6. Typical spring day, Southwest, High Renewables Case

Notwithstanding these similarities, the Southwest does have an important

distinguishing characteristic from California: its reliance on coal generation for a

large share of its installed capacity. Historically, the coal fleet in this region has

operated largely in a baseload capacity, ramping and cycling with limited

frequency. The ability to operate the coal fleet more flexibly during high

renewable output days is a key strategy to facilitate renewable integration in

the Southwest and in other regions that rely on coal generation today.

The third region that experiences significant quantities of renewable

curtailment in the High Renewables Case is the Northwest; though oversupply is

its primary cause, the nature of the oversupply is entirely unique to that region.

The Northwest is characterized by its large hydroelectric fleet, and during the

spring runoff season, during average and wet hydro years, the Northwest hydro


P a g e | xviii |

system is capable of meeting most or all of the Northwest’s energy needs on a

day‐to‐day basis. Adding wind generation to that system during that time of

year results in an excess supply of zero‐marginal‐cost generation. Such

oversupply events have already been experienced in parts of the Northwest in

2010‐2012 when surplus hydro conditions, combined with dispatch inflexibility,

led to wind curtailment.7 With the addition of large amounts of wind generation

in the High Renewables Case, the oversupply during the spring runoff is

intensified. An example of such a day is shown in Figure 7.

Figure 7. Typical spring day, Northwest, High Renewables Case

7 In recognition of this phenomenon, the Bonneville Power Administration (BPA) has developed an “Oversupply Management Protocol” to allow for prudent management of generation resources during oversupply conditions.

Executive Summary

P a g e | xix |© 2015 Energy and Environmental Economics, Inc.

Spring days are chosen to highlight flexibility challenges because the spring

months (March through June) represent the most challenging periods for

renewable integration throughout the West. The combination of relatively low

loads, high hydro system output, and high renewable output results in a

concentration of oversupply during this period of the year. Figure 8 illustrates

the diurnal and seasonal trends in renewable curtailment for these three

regions through heat maps that display the average amount of renewable

curtailment in each month‐hour. The similarities between California and the

Southwest are immediately evident: throughout the year, curtailment is

concentrated in the middle of the day, driven by the solar PV oversupply in each

region. This oversupply is greatest in the spring and smallest in the summer,

when each region’s loads are highest due to high cooling loads and can absorb

higher quantities of solar PV output. The seasonal trend in the Northwest, by

contrast, is more closely linked to the spring runoff from the hydro system and

its interaction with a wind‐heavy renewable fleet, which leads to renewable

curtailment at all hours in the spring but is most concentrated at night when

wind output is high and loads are low.


P a g e | xx |

Figure 8. Seasonal and diurnal patterns of renewable curtailment, High Renewables Case

The dramatic increase in curtailment observed in these three regions between

the Common Case and the High Renewables Case is indicative of a phenomenon

that is highly nonlinear. In the absence of mitigation measures, curtailment

grows at an increasing rate as renewable penetration climbs. For example, the

marginal curtailment for a new solar PV resource (i.e. the next solar resource

(a) Northwest1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Scale (MW)

Jan 4,000

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec 0

(b) California1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Scale (MW)

Jan 16,000

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec 0

(c) Southwest1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Scale (MW)

Jan 7,000

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec 0

8,000

2,000

3,500

Executive Summary

P a g e | xxi |© 2015 Energy and Environmental Economics, Inc.

built on top of the portfolios described here) in California is between 50‐60%—

that is, in excess of half of its available energy would not be able to be delivered

to the California system—a figure substantially higher than the average

curtailment of 9%.

In contrast to California and the Northwest and Southwest regions, curtailment

is observed in much lower quantities in both the Basin and Rocky Mountain

regions. What distinguishes the Basin and Rocky Mountain regions from

California and the Southwest—all four fleets rely primarily on dispatchable

thermal resources to serve net load—is the amount of diversity assumed in their

renewable portfolios. Whereas the concentration of solar PV during the middle

of the day drives the oversupply phenomenon in California and the Southwest,

production is far less concentrated during specific periods of the year in the

Basin and Rocky Mountain regions. In the Basin, this is a result of technological

diversity that combines geothermal, solar and wind resources; in the Rocky

Mountains, it is the result of a portfolio comprising primarily wind resources

whose scale and geographic dispersion result in production spread throughout

the year.

While curtailment is observed in limited quantities in these regions, both the

Rocky Mountains and the Basin region show dramatic changes in how the

generation fleets operate to balance higher penetrations of renewable

generation. A key driver of the magnitude of operational challenges in these

two regions is the extent of the ability of thermal generators to ramp and cycle.

Historically, coal generators have operated as baseload resources, running at

high capacity factors throughout the year. Higher penetrations of renewable

generation will exert pressure on operators to use these units more flexibly than


P a g e | xxii |

they have historically; however, there are technical, economic, and institutional

factors that might limit their ability to do so. The amount of flexibility that can

be harvested from the aging coal fleet is uncertain and deserves further

attention, but its implications for renewable integration are clear: an electric

system with coal generators whose flexibility is limited in day‐to‐day operations

will experience renewable curtailment in larger quantities and more frequently

than one in which coal generation operates flexibly. This tradeoff is illustrated

for both the Basin and Rocky Mountain regions in Figure 9.

Figure 9. Typical spring days, Basin and Rocky Mountains, High Renewables Case (with and without limits on coal flexibility)

Executive Summary

P a g e | xxiii |© 2015 Energy and Environmental Economics, Inc.

Enabling Strategies for Renewable Integration

This study treats renewable curtailment as the “default solution” to flexibility

challenges as it represents a last recourse for system operators—prior to

shedding load—once the flexibility of the existing traditional dispatchable

system has been exhausted. However, there are many options for alternative

strategies that, if identified and deployed in advance, can provide operators

with additional flexibility and mitigate the need to curtail renewable generation.

These “solutions” include improvements in scheduling and dispatch,

investments in new flexible generation, and new demand‐side programs.

Investigating the full potential list of integration solutions is beyond the scope of

this study; the solutions chosen for study herein are intended to help illustrate

the types of attributes that flexibility planners may wish to consider. This study

considers three specific measures to facilitate renewable integration:

Increased regional coordination;

Investments in energy storage technologies; and

Investments in flexible gas generation.

The first among these—improving regional coordination in operations—

represents an institutional improvement to facilitate renewable integration.

Historically, the balkanized Western Interconnection has operated largely

according to long‐term contractual agreements and through bilateral power

exchange. “Regional coordination” could represent a step as incremental as

improving upon existing scheduling processes or as comprehensive as the

consolidation of the Western Interconnection under a single centralized

operator. This study does not presume the mechanism through which


P a g e | xxiv |

coordination occurs, but demonstrates the value of achieving full coordination

by contrasting a future in which interregional exchange is limited to the

historical range with one in which it is limited by the ratings of interregional

transmission paths, allowing fuller utilization of load and resource diversity.

Relaxing the constraint on interregional exchange to allow the use of the

transmission system to its physical limits results in a reduction of renewable

curtailment from 6.4% to 3.0% (see Table 2). A significant share of this value is

derived from regions that, facing an oversupply condition that might otherwise

require renewable curtailment, are able to use the full capability of the

transmission system to find an alternative market for their power. This impact is

shown in Figure 10. The reduction of curtailment is largest in California—which

is both electrically and institutionally close to a market in the Northwest that

can accommodate midday exports outside of the spring season—and in the

Northwest—whose nighttime oversupply finds a destination in both California

and the Southwest markets. Notably, the observed impact on the Southwest is

lower than in the prior two regions, as connections with the Basin and Rocky

Mountains provide it with limited export markets and California’s frequent

simultaneous oversupply makes it an impractical market for solar surplus.

Table 2. Comparison of regional renewable curtailment with historical and physical intertie limits imposed, High Renewables Case

Scenario Basin California Northwest Rockies Southwest

Historical Intertie Limits 0.4% 8.6% 6.1% 0.1% 7.3%

Physical Intertie Limits 0.5% 3.1% 1.6% 0.5% 6.0%

Difference +0.1% ‐5.6% ‐4.5% ‐0.0% ‐1.3%

Executive Summary

P a g e | xxv |© 2015 Energy and Environmental Economics, Inc.

Figure 10. Impact of increased regional coordination upon seasonal curtailment patterns.

(a) Basin1 24 1 24 Scale (MW)

Jan JanFeb FebMar MarApr AprMay MayJun JunJul JulAug AugSep SepOct OctNov NovDec Dec

(b) California1 24 1 24 Scale (MW)


(c) Northwest1 24 1 24 Scale (MW)


(d) Rocky Mountains1 24 1 24 Scale (MW)


(e) Southwest1 24 1 24 Scale (MW)


Historical Intertie Limits Physical Intertie Limits

0

7,000

0

4,000

0

200

3,000

2,000

1,000

150

100

50

5,250

3,500

1,750

200

0

15,000

0

150

100

50

11,250

7,500

3,750


P a g e | xxvi |

While desirable from a technical and economic perspective, harvesting the full

diversity of the Western Interconnection through coordination of operations to

integrate renewable generation presents a significant institutional challenge.

Individual utilities, balancing authorities, and planning entities may be justified

in examining potential investments or demand‐side programs to provide greater

local or regional operational flexibility. To examine the potential flexibility

benefits of investments in new flexible resources, 6,000 MW of energy storage8

and new fast‐starting, fast‐ramping gas CCGTs are added to the High

Renewables Case in separate scenarios to quantify their impact on operations.9

The impact of each additional resource on renewable curtailment is summarized

in Table 3.

Table 3. Impacts of new investments on regional renewable curtailment, High Renewables Case.

Scenario Basin California Northwest Rockies Southwest

Reference Grid 0.4% 8.7% 5.6% 0.6% 7.3%

+6,000 MW Storage (2hr) 0.4% 7.2% 5.7% 0.6% 6.2%

+6,000 MW Storage (6hr) 0.4% 5.8% 5.8% 0.6% 5.1%

+6,000 MW Storage (12hr) 0.4% 5.8% 5.7% 0.6% 5.1%

+6,000 MW Flex CCGT 0.4% 8.7% 5.6% 0.6% 7.3%

The primary result this analysis indicates is that the addition of new

“downward” flexibility (e.g. the ability to charge energy storage) provides

substantially greater benefits than the addition of new “upward” flexibility (e.g.

8 Multiple durations of energy storage resources were modeled, including 2‐hr, 6‐hr, and 12‐hr. 9 In each scenario, 4,000 MW of the candidate resource was located in California and 1,000 MW was located in both the Northwest and the Southwest. The choice of where to add storage was made on the basis of where the largest flexibility challenges were identified.

Executive Summary

P a g e | xxvii |© 2015 Energy and Environmental Economics, Inc.

the ability to ramp from Pmin to Pmax very quickly), as it expands the net load

range across which a system can operate. This analysis indicates several findings

on the impacts of new flexibility investments:

Energy storage provides a clear benefit through curtailment mitigation

in the Southwest and California. This is largely due to the fact that each

region’s diurnal solar oversupply allows storage resources to cycle

almost daily, charging during the middle of the day during curtailment

hours and discharging in the early morning and/or evening to help meet

solar‐driven net load ramps.

The value of energy storage in the Northwest is limited by comparison.

Here, where oversupply events during the spring runoff persist much

longer, often throughout the day, there is a much more limited

opportunity to shift generation within the day.10 Also, the Northwest

already has significant intra‐day energy storage capability through its

existing system of hydroelectric resources.

In all regions where new flexible gas CCGT capacity is added, the impact

is minimal. While further study is needed, it appears that all regions

have sufficient upward ramping capability in the base system. These

new, efficient gas resources may reduce system cost due to their

efficiency, but they do little to reduce curtailment because they do not

appreciably improve the system‐wide ratio of minimum to maximum

generation capability.

10 Note that the REFLEX approach of using day draws likely understates the value of storage in the Northwest, as it

does not capture value that could be provided through inter‐day charging and discharging patterns.


P a g e | xxviii

While the general findings of these cases help to illustrate the attributes of new

flexible resources that provide significant benefits, the limitations of their

applicability must also be understood:

One of the underlying assumptions of the regional approach is that each

constituent utility and balancing authority of a region makes its

resources fully available for optimal dispatch within that region.

However, this assumption is optimistic for today’s electric system—one

that is dominated by bilateral transactions—and findings that apply to a

wholly optimized region may not be applicable to individual entities

within it.

The starting case to which the flexible resources are added assumes a

significant degree of flexibility in the existing thermal system.

Production simulation models typically do not consider increased O&M

costs incurred as units are asked to cycle more frequently. As the

sensitivity on coal flexibility demonstrates, there is value to a system

that can operate its thermal units flexibly (i.e. can reduce their output

to very low levels or turn off), and to the extent that technical,

economic, or institutional factors prevent the coal fleet from operating

as this study assumes, additional flexible thermal generation could have

more value than this study identifies.

A critical qualifier for this analysis is that each regional fleet already

meets traditional resource adequacy needs prior to the addition of

incremental flexible capacity. While the value of flexibility alone may

not be enough to justify procuring a specific resource, entities should

carefully consider the benefits that additional flexibility could provide

when new system capacity is needed due to load growth, retirement of

aging coal generation, or other factors.

Executive Summary

P a g e | xxix |© 2015 Energy and Environmental Economics, Inc.

Implications and Next Steps

The technical findings and conclusions reached through this study have a

number of implications that are relevant for regulators and policymakers

seeking to enable higher penetrations of renewable generation on the system

and to ease the associated challenges.

1. The analysis conducted in this study identifies no technical barriers to the

achievement penetrations of renewable generation of up to 40% of total

supply in the Western Interconnection.

In both the Common Case and the High Renewables Case, the flexibility

assessment demonstrates capability of the electric system of the Western

Interconnection to serve loads across a diverse range of system conditions.

Further, no “need” for additional flexible capacity beyond existing and planned

resources is identified in either case. Examination of issues relating to stability,

adequacy of frequency response, and the potential impact of major

contingencies is beyond the scope of this study; however, other studies to

examine these topics are have come to similar conclusions regarding the

technical feasibility of integrating high penetrations of renewable generation.11

What does distinguish the High Renewables Case from the Common Case is the

significant quantity of renewable curtailment observed. At relatively low

penetrations, renewable generation can be integrated into the system

11 See the Western Wind and Solar Integration Study – Phase 3 (NREL & GE), available at:

http://www.nrel.gov/electricity/transmission/western‐wind‐3.html


P a g e | xxx |

effectively with limited curtailment; however, once a region’s penetration

surpasses a certain threshold, curtailment occurs with increasing frequency.

Renewable curtailment is characteristic of all of the High Renewables scenarios

investigated: Figure 11 summarizes the range of curtailment frequency across

all of these scenarios.

Figure 11. Observed range of hourly curtailment frequency across all High Renewables scenarios in each region.

2. Routine, automated renewable curtailment is a fundamental necessity to

electric systems at high renewable penetrations, as it provides operators with

a relief valve to manage net load conditions to ensure a system can be

operated reliably at increasing penetrations.

Historically, the idea of “curtailment” has had negative connotations, but at high

penetrations of renewable generation, it is a fundamental necessity.

Executive Summary

P a g e | xxxi |© 2015 Energy and Environmental Economics, Inc.

Notwithstanding its cost to ratepayers, renewable curtailment provides

operators with a tool with which to manage the net load to ensure reliable

service when the flexibility of all other existing resources has been exhausted; in

this respect, it serves as the “default solution” for a system constrained on

flexibility.

The role of curtailment in the operations of an electric system at high renewable

penetrations is multifaceted. In addition to allowing operators to mitigate

oversupply events when generation would otherwise exceed loads, renewable

curtailment can be used to mitigate the size of net load ramps across one or

more hours, to adjust for forecast error between day‐ahead and other

scheduling processes, and to substitute for traditional flexibility reserve services

(including regulation).

Because of the crucial role of renewable curtailment in operating electric

systems at high penetrations of renewables, ensuring that curtailment is

available and can be used efficiently in day‐to‐day operations requires a number

of steps:

Market structures and scheduling processes must be organized to

allow participation by renewable generators. Within organized

markets, this means ensuring that utilities can submit bids into the

market on behalf of renewable generators that reflect the opportunity

cost of curtailing these resources as well as ensuring that renewable

plants are not excessively penalized for deviations from their schedules

due to forecast errors. In environments in which vertically integrated

utilities or another type of scheduling coordinator is responsible for

determining system dispatch, the operator must begin to consider the


P a g e | xxxii

role of renewable curtailment in scheduling and dispatch decisions for

both renewable and conventional resources.

Contracts between utilities and renewable facilities must be

structured to allow for economic curtailment. Historically, many power

purchase agreements have been set up to pay renewables for the

generation that they produce and have included provisions limiting

curtailment under the premise that limiting risk and ensuring an

adequate revenue stream to the project are necessary to secure

reasonable financing. Compensated curtailment, under which

developers are paid a PPA price both for generation that is delivered to

the system as well for estimated generation that is curtailed, would be

one means of achieving this goal.

To the extent that additional institutional barriers that would limit operators

from effectively dispatching renewable resources exist, these barriers must also

be addressed to allow for renewable integration at higher penetrations.

3. Another key step to enabling reliable and efficient operations under high

penetrations is ensuring operators fully understand the conditions and

circumstances under which renewable curtailment is necessary or desirable.

In some instances—namely, during oversupply conditions—the need to curtail is

relatively intuitive; however, in other instances, the important role of

curtailment may not be so obvious. For example, an operator faced with a

choice between keeping a specific coal unit online and curtailing renewables or

decommitting that coal unit to allow additional renewable generation should

make that decision with knowledge of the confidence in the net load forecast as

well as an understanding of the consequences of possible forecast errors.

Similarly, an operator anticipating a large upward net load ramp may decide to

Executive Summary

P a g e | xxxiii |© 2015 Energy and Environmental Economics, Inc.

curtail renewable generation prospectively to spread the ramp across a longer

duration if the ramp rates of conventional dispatchable units are limited.

Additional work is necessary to identify such operating practices and conditions

in which renewable curtailment may be necessary outside of oversupply

conditions to ensure reliable service.

The role of operating reserves at avoiding unserved energy under unexpected

upward ramping events must also be considered. Resources under governor

response or Automated Generation Control (AGC) respond quickly to small

deviations in net load. Contingency reserves (spinning and supplemental or

“non‐spin” reserves) are used to manage large disturbances such as the sudden

loss of a generator or transmission line. Additional categories of reserve

products—for instance, “load following” or “flexibility” reserves—have been

contemplated at higher renewable penetration, but have not yet been

formalized. How these reserves are deployed will impact the magnitude of

challenges encountered at higher renewable penetrations.

4. The consequences of extended periods of negative pricing must be

examined and understood.

Historically, the centralized markets and bilateral exchanges of the Western

Interconnection have, for the most part, followed the variable costs of

producing power—most often the costs of fuel and O&M for coal and gas

plants. In a future in which renewable curtailment becomes routine, forcing

utilities to compete to deliver renewable generation to the loads to comply with

RPS targets, the dynamics of wholesale markets will change dramatically. How

the dynamics of negative pricing ultimately play out remains a major


P a g e | xxxiv

uncertainty; nonetheless, with frequent low or negative prices in a high

renewables future, utilities, other market participants, and regulators will be

confronted by a host of new questions:

How should generators that provide other services to the system during

periods of low negative prices be compensated?

How can the proper signal for investment in generation resources be

provided as frequent negative prices further erode margins in energy

markets?

Do negative prices create new issues for loads, who, rather than paying

for power from the wholesale market during periods of curtailment,

would be paid to consume?

At what point does the prevalence of negative prices lead to new policy

mechanisms other than production quotas to promote the development

of new renewable energy?

How should future retail tariffs be designed to balance considerations of

equity and cost causation with the radical changes in wholesale market

signals?

These and other questions will require consideration as penetrations of

renewables continue to increase.

5. While renewable curtailment is identified as the predominant challenge in

operations at high renewable penetrations, its magnitude can be mitigated

through efficient coordination of operations throughout the Western

Interconnection.

Executive Summary

P a g e | xxxv |© 2015 Energy and Environmental Economics, Inc.

The balkanization of operations today presents an institutional barrier to

efficient renewable integration; by allowing full utilization of the natural

diversity of loads and resources throughout the Western Interconnection,

regional coordination offers a low‐hanging fruit to mitigating integration

challenges. A number of studies have identified the significant operational

benefits that can be achieved through balancing authority consolidation, a

conclusion that is supported by the reduction in renewable curtailment at high

penetrations identified in this study.

6. Many supply‐ and demand‐side solutions merit further investigation to

understand their possible roles in a high renewable penetration electric

system.

This study examines a select few of the multitude of possible supply‐ and

demand‐side portfolio measures available to utilities to illustrate how different

attributes do (or do not) provide value to electric systems at high penetrations

of renewable generation. The solutions examined within this study illustrate

how different “types” of flexibility impact a system to differing degrees:

whereas storage effectively mitigates renewable curtailment through its ability

to charge during periods of surplus, fast‐ramping flexible gas resources have a

comparatively limited impact on operations, displacing less efficient gas

generation resources but effecting minimal changes in curtailment.

7. The ability of renewable curtailment to serve as an “avoided cost” of

flexibility points to an economic decision‐making framework through which

entities in the Western Interconnection can evaluate potential investments in

flexibility and ultimately rationalize procurement decisions.


P a g e | xxxvi

As the need for operational flexibility has grown, a number of efforts have

explored whether additional planning standards—analogous to those used for

resource adequacy today—are necessary to ensure that when the operating day

comes, the generation fleet is sufficiently flexible to do serve load reliably. As

this study demonstrates, so long as (1) the generation fleet is capable of

meeting extreme peak demands, and (2) the operator can use curtailment as a

relief valve for flexibility constraints, the operator can preferentially dispatch

the system to avoid unserved energy. Thus, the consequence of a non‐

renewable fleet whose flexibility is inadequate to balance net load is renewable

curtailment, whose implied cost is orders of magnitude smaller than the cost of

unserved energy. In this respect, the determination of flexibility adequacy is

entirely different from resource adequacy: for resource adequacy, conservative

planning standards are justified on the basis of ensuring that costly outages are

experienced exceedingly rarely; for flexibility adequacy, the appropriate amount

of flexibility for a generation system is instead an economic balance between

the costs of “inadequacy” (renewable curtailment) and the costs of procuring

additional flexibility.

Because renewable curtailment can serve as an “avoided cost” of flexibility, the

question of “flexibility adequacy” is economic, rather than technical. Renewable

curtailment imposes a cost upon ratepayers, reflected in this study by the idea

of the “replacement cost,” and, to the extent it can be reduced through

investments in flexibility, its reduction provides benefits to ratepayers. At the

same time, designing and investing in an electric system that is capable of

delivering all renewable generation to loads at high penetrations is, itself, cost‐

prohibitive. Between these two extremes is a point at which the costs of some

new investments or programs that provide flexibility may be justified by the

Executive Summary

P a g e | xxxvii © 2015 Energy and Environmental Economics, Inc.

curtailment they avoid, but the cost of further investments would exceed the

benefits. This idea is illustrated in Figure 12, which shows the tradeoff between

the costs of renewable curtailment with the costs of a possible theoretical

measure undertaken to avoid it. This study provides both an example of the

type of analytical exercise that could be performed to quantify the operational

benefits of flexibility solutions as well as a survey of the analytical

considerations and tradeoffs that must be made in undertaking such an

exercise.

Figure 12. Illustration of an economic framework for flexibility investment.

While not performed in the context of this study, this type of economic

assessment of flexibility solutions to support renewables integration will depend

on rigorous modeling of system operations combined with accurate

representation of the costs and non‐operational benefits of various solutions.

The specific types of investments to enable renewable integration that are


P a g e | xxxviii

found necessary will vary from one jurisdiction to the next, but the overarching

framework through which those necessary investments are identified may be

consistent. Implementation of such an economic framework for decision‐making

for flexibility will foster the transition to high renewable penetration, enabling

the achievement of policy goals and decarbonization while mitigating the

ultimate impacts of those changes to the quality and cost of service received by

ratepayers.

western interconnection flexibility assessment · 2016-01-14 · 101 montgomery street, suite 1600...

Documents